takes care of the government contracts bidding
Country
Project Duration
Team
The client is an advanced AI project that automates the entire process of searching for and bidding on government contracts.
The system simplifies and streamlines the contract bidding process, making it more efficient for companies to compete for government projects.
Searching for relevant government contracts across multiple websites.
Utilizing AI to accurately identify suitable contracts for bidding.
This mechanism is needed for comparing client prices with those of competitors.
Creating and submitting bids automatically with AI.
Improving the response quality from Large Language Models (LLMs) like GPT-3.5, GPT-4, GPT-4o, and Claude.
Scraping content from various sites and Google Search results efficiently and reliably.
Processing multiple documents simultaneously to minimize wait times.
Managing multiple resource-intensive tasks and a large codebase simultaneously.
Selected and integrated AI algorithms to recognize suitable contracts and identify bidding opportunities.
Automated the process of comparing client prices with competitors to ensure competitive bids.
Built an AI system to create detailed bids and automatically send them via email.
Engineered different prompts and worked with various LLM models (GPT-3.5, GPT-4, GPT-4о, Claude), testing each multiple times with different contexts to achieve the most appropriate results for the domain of bidding.
Utilized modern tools and libraries like Cheerio, ZenRows, and SerpAPI for efficient and fault-tolerant scraping, enriching our database and business logic.
Created true parallelization of processing in Node.js by utilizing Node.js Workers (Threads), reducing the overall wait time for results by more than half.
Divided the project into two microservices and moved resource-intensive tasks to a separate service, ensuring task delivery with RabbitMQ, which minimized the impact on REST API performance and improved overall backend efficiency.
Developed a comprehensive web application from scratch, including both design and functionality.
Integrated sophisticated AI algorithms for contract recognition and bid generation.
Improved data processing through parallelization, efficient scraping, and fault-tolerant mechanisms.
Implemented a microservices architecture to handle heavy-weight tasks and large codebase efficiently.
Enhanced the quality of responses from various LLMs for better domain-specific results.
React JS
Nest JS
Python
ZenRows
SerpAPI
Node.js
RabbitMQ
Our team successfully developed the auto-bidding platform from the ground up, automating the entire government contract bidding process.
Despite challenges in data retrieval, AI integration, and handling large-scale tasks, we delivered a robust and efficient system. This ongoing project continues to evolve, and the client is extremely satisfied with the results, which have significantly accelerated and improved the bidding process.
